Research And Implementation On Shortwave Infrared Image Preprocessing For Push Broom Hyperspectral Imager

The shortwave infrared image data collected by push boom hyperspectral imager has some characteristics, such as multi band, high spectral resolution and combination of image and spectrum and so on. The imager has strong ability to identify land cover and can obtain the approximate continuous spectral information. The pre-processing of Shortwave infrared image is the most important precondition for the quantitative analysis of image information and the application of the infrared image.Therefore, this thesis studied the pre-processing technologies of shortwave infrared image collected by push boom hyperspectral imager. Based on this, we designed and implemented the pre-processing software for shortwave infrared image data. The main works of this thesis can be summarized as follows:(1) The pre-processing techniques of shortwave infrared image were briefly introduced, including quantitative processing technology, radiation correction technology and non-uniformity correction technology.(2) Many characteristics including radiation transmission, integration time parameters, dark current and noise of problem hyperspectral imager were analyzed in detail, which provided the theoretical basis for the subsequent data processing. Based on the mathematic model, we finished the data fitting and the comparison.(3) In order to improve the quality of the shortwave infrared image, we focused on the radiation correction technology, noise filtering technology and non-uniformity correction technology. The algorithm implementation process and the experimental results were given in this thesis, the results were also analyzed. Relying on the thorough study of non-uniformity correction processing algorithm for shortwave infrared image, we proposed an based on homogeneous feature algorithm for non-uniformity issues. The mean of Digital Number values of the single uniform objects was regarded as the theoretical value. We calculated the non-uniformity correction coefficients pixel by pixel. Finally, we completed the non-uniformity correction for images and verified that the algorithm can effectively improve the shortwave infrared images’ quality.(4) In the Windows system, by using Microsoft Foundation Class Library (MFC) under C++programming environment, we designed a data pre-processing system software which can realize data pre-processing for shortwave infrared images.